Contents lists available at ScienceDirect

Fisheries Research

journal homepage: www.elsevier.com/locate/fishres

Using harmonized historical catch data to infer the expansion of global tunafisheries

Angie Coultera,⁎, Tim Cashiona,b, Andrés M. Cisneros-Montemayorc, Sarah Popova, Gordon Tsuia,Frédéric Le Manachd, Laurenne Schillere, Maria Lourdes D. Palomaresa, Dirk Zellerf,Daniel Paulya

a Sea Around Us, Global Fisheries Cluster, Institute for the Oceans and Fisheries, University of British Columbia, Vancouver, Canadab Fisheries Economics Research Unit, Global Fisheries Cluster, Institute for the Oceans and Fisheries, University of British Columbia, Vancouver, CanadacNereus Program, Institute for the Oceans and Fisheries, University of British Columbia, Vancouver, Canadad BLOOM Association, Paris, FranceeMarine Affairs Program, Dalhousie University, Halifax, NS, Canadaf Sea Around Us – Indian Ocean, School of Biological Sciences, University of Western Australia, Perth, Australia

A R T I C L E I N F O

Handled by George A. Rose

Keywords:Large pelagic fishesBillfishesFisheries managementBycatchDiscardsRFMO

A B S T R A C T

Despite worldwide demand for tuna products and considerable conservation interest by civil society, no singleglobal dataset exists capturing the spatial extent of all catches from fisheries for large pelagic species across allocean basins. Efforts to spatially quantify the historical catch of global tuna fisheries have been restricted to thefew taxa of major economic interest, creating a truncated view of the true extent of the fisheries for tuna andother large pelagic fishes. Individual Regional Fisheries Management Organizations (RFMOs) have given varyingdegrees of attention to minor taxa and non-target species only in more recent years. Here, we compiled andharmonized public datasets of nominal landed catches, as well as spatial data on reported catches of large pelagictaxa reported for the industrial tuna and large pelagic fisheries by tuna RFMOs for the last 60+ years.Furthermore, we provide a preliminary estimate of marine finfishes discarded by these fisheries. We spatializedthese data to create a publicly available, comprehensive dataset presenting the historical reported landed catchesplus preliminary discards of these species in space for 1950–2016. Our findings suggest that current publicreporting efforts are insufficient to fully and transparently document the global historical extent of fisheries fortuna and other large pelagic fishes. Further harmonization of our findings with data from small-scale tunafisheries could contribute to a fuller picture of global tuna and large pelagic fisheries.

1. Introduction

What are generally called “tuna fisheries”, i.e., fisheries for largepelagic tuna, billfishes and pelagic sharks are some of the oldest fish-eries in the world, with evidence that humans had pelagic fishingcapabilities over 40,000 years ago (O’Connor et al., 2011). Coastal ar-tisanal fisheries for large pelagic species have existed for millennia intropical and sub-tropical areas, while industrial efforts developed overthe last century. Japan was the first country to foray into industrialfisheries for tuna in the 1920s, investing in bait boat operations in thePacific Islands (Gillett, 2007). With an increasing demand for cannedseafood, the industrial fishing effort intensified after World War II(Miyake et al., 2004). Improvements in vessel technology and freezercapabilities allowed these industrial tuna fisheries to rapidly expand,

with fleets operating across virtually all of the Atlantic, Indian andPacific Oceans since the 1980s (Majkowski, 2007).

Demand for tuna and tuna-like species is at an all-time high, withrecord reported landings of 7.7 million t in 2014 (FAO, 2016), afterwhich it levelled off to around 7.5 million t·year−1 since (FAO, 2018).Tuna fisheries can be highly profitable, partly due to subsidies thatoffset their often high operational costs (Lam et al., 2011; Sumailaet al., 2016), although economic and cost challenges remain (Sala et al.,2018), as do concerns about resource-use equity (Sumaila et al., 2015)and relevance to food security (Schiller et al., 2018). There are cur-rently five Regional Fisheries Management Organizations (RFMOs) re-sponsible for ensuring long-term sustainability of the stocks under theirjurisdiction: the Commission for the Conservation of Southern BluefinTuna (CCSBT), the Indian Ocean Tuna Commission (IOTC), the Inter-

https://doi.org/10.1016/j.fishres.2019.105379Received 2 December 2018; Received in revised form 12 September 2019; Accepted 13 September 2019

⁎ Corresponding author.E-mail address: a.coulter@alumni.ubc.ca (A. Coulter).

Fisheries Research 221 (2020) 105379

0165-7836/ © 2019 Elsevier B.V. All rights reserved.

T

American Tropical Tuna Commission (IATTC), the International Com-mission for the Conservation of Atlantic Tunas (ICCAT), and the Wes-tern and Central Pacific Fisheries Commission (WCPFC).

While the scope and research capacity of each RFMO differs, allfacilitate data collection and stock assessment in their managementarea (Fig. 1) and set regulations on fishing activity for member coun-tries. However, there continue to be instances of considerable over-fishing of high-seas fish stocks, and all RFMOs have been challenged ontheir ability to successfully manage stocks and fisheries (Cullis-Suzukiand Pauly, 2010, 2016) as well as protect non-target species impactedby fishing fleets (Juan-Jordá et al., 2018).

Despite the high economic value of tuna landings (FAO, 2016) andthe considerable interest by the general public in the conservation andsustainable management of large pelagic species, there is, to date, nosingle, combined global dataset presenting all catches of species caughtby these fisheries over time and space. The Food and Agriculture Or-ganization of the United Nations (FAO) produces the ‘Atlas of Tuna andBillfish Catches’ (referred to here as the ‘FAO Tuna Atlas‘), which coversonly the catch of 12 species of tuna and billfishes. The FAO Tuna Atlas(FAO, 2017) presents data only for albacore tuna (Thunnus alalunga),Atlantic bluefin tuna (T. thynnus), Atlantic white marlin (Kajikia albida),bigeye tuna (T. obesus), black marlin (Istiompax indica), blue marlin(Makaira nigricans), Pacific bluefin tuna (T. orientalis), skipjack tuna(Katsuwonus pelamis), southern bluefin tuna (T. maccoyii), stripedmarlin (K. audax), swordfish (Xiphias gladius), and yellowfin tuna (T.albacares). While this dataset encompasses the major target species, itdoes not account for the landings of other species of commercial andconservation interest (e.g., sharks), which are reported by RFMOs, nordoes it address discarded catch (Zeller et al., 2018).

Additional species-specific information on landed catches is avail-able from national fisheries statistics presented by FAO on behalf ofmember countries (Garibaldi, 2012), but these data explicitly do notaccount for discarded bycatch, nor provide details on gear types used orfisheries sectors, nor include detailed spatial information (FAO, 2019).Addressing this data gap requires a synthesis and harmonization of allspatial data with all nominal catch data reported by each of the tunaRFMOs, as well as the addition of data on discards, which will allow fora better understanding of the full scope of global fisheries for largepelagics over the last six decades. A recent initiative by the Institut de

Recherche pour le Développement (IRD) and the FAO has aimed to har-monize global tuna catches across space and time for all RFMOs(Taconet et al., 2017). However, while a great improvement from iso-lated RFMO data repositories, the dataset of Taconet et al. (2017) doesnot spatialize all nominal catches or estimate discards for all regions,but only for subsets of data as provided by the RFMOs. Therefore, webuild on these excellent, initial initiatives, and the preliminary workdescribed in Le Manach et al. (2016), to develop and present a har-monized, comprehensive and spatialized global dataset of historicallarge pelagics catch data from 1950 to 2016. It differs from other stu-dies by spatializing all nominal reported landings, as well as pre-liminary estimates of discards. Results provide a global overview oftuna catches over the last sixty years, and allow for analysis of large-scale patterns in space and time that can combine with ongoing re-gional data collection to inform better long-term policies.

2. Methods

2.1. Source data

We assembled public domain datasets from all five tuna RFMOs(Table 1), which include catch data reported by member countriesfishing in the corresponding management areas. For each RFMO, thisinformation is provided in two separate datasets. Firstly, the ‘nominal’catch dataset represents all reported catches of species within eachRFMO’s purview in a given year, and includes information on thefishing country, gear (or gear type), taxon, time (year), ocean area (forRFMOs spanning more than one, e.g., North and South Pacific), andweight of catch (here deemed to represent whole, wet weight). Sec-ondly, the ‘spatial’ catch dataset is a subset of nominal data (i.e., not allyearly catches are included for most countries) that includes georefer-enced information on the spatial location of catch. The ‘spatial’ datasetfrom the WCPFC is an exception, as it does not report fishing countryfor confidentiality reasons.

The aim of our methodology was to harmonize and spatialize theseseparate datasets using a rule-based, step-wise approach to match allnominal catch data to corresponding spatial data, geographically re-fined in what we call ‘tuna blocks’, i.e., the spatial data reporting blocksranging from 1° x 1° to 20° x 20° as used by RFMOs for spatial reporting.

Fig. 1. Areas of responsibility for each of the tuna Regional Fisheries Management Organizations (RFMO). ICCAT: International Commission for the Conservation ofAtlantic Tunas; IOTC: Indian Ocean Tuna Commission; CCSBT: Commission for the Conservation of Southern Bluefin Tuna; WCPFC: Western and Central PacificFisheries Commission; IATTC: Inter-American Tropical Tuna Commission.

A. Coulter, et al. Fisheries Research 221 (2020) 105379

2

As a second goal, we added preliminary estimates of discards to allreported catches.

2.2. Spatialization method

For each ocean basin, the nominal catch data were spatialized totuna blocks according to the reported proportions in the spatial data.The ideal, best case spatialization scenario assigned a given nominalcatch entry to tuna blocks using only spatial catch data with an exactmatch for all data categories (i.e., where both datasets contain catchrecords for the same country, year, taxon, and gear). For example, if thenominal dataset for a given RFMO included a record of France catching10 t of bigeye tuna using longline in 1983, and the spatial dataset in-cluded records of France catching 8 t of bigeye tuna using longline in1983, with 4 t caught in spatial tuna block A and 4 t in block B, the 10 treported in the nominal catch would be assigned as 5 t in each of thetwo tuna blocks. Thus, the total catch reported in the nominal data(which in theory represents the total reported catch in a given year)would be unchanged, but allocated to its most likely catch locationsbased on available spatial records. We thus assumed that the reported,spatial subset of data as nearly as possible represented the most likelytrue spatial extent of all catches with the given data categories.

This matching of the nominal and spatial catch data records wasrepeated over a series of successive refinements. After all full data ca-tegory matches were spatially allocated, we focused on matching allnominal catch records with less than perfect data category match withspatial data for one fishing country (flag state) at a time. With eachsuccessive matching iteration, the year span of spatial data being usedto match nominal data records were relaxed to intervals of +/-2 andthen +/- 5 years. After attempting to match all nominal and spatialrecord categories over all year ranges, the procedure was re-run usingsuccessively less-precise data category matches and including cate-gories for groups of gears and groups of species. Each successivematching step was used before expanding the year ranges further to 10,20, and 35 years. While these year ranges were much larger than wouldbe ideal, they were necessary to capture the limited spatial data forsome RFMOs. This applied especially to the IOTC, which has nominaldata only from 1952 and limited spatial data from 1952 to 1966, butwith better spatial coverage from 1967 to present. These wider yearranges were only utilized to spatialize data if no other records for closeryears were available. Thus, our final spatialized dataset is likely morebroadly assigned for earlier years.

After each successive refinement, the matched and unmatched re-cords were stored separately, so that at each new refinement, only theprevious step’s unmatched records were used. The end-result was acatch database where all nominal catch records were spatialized to theexisting global spatial catch records by tuna blocks of various sizes(Table 1) and containing all original nominal records. Thus, the finalspatial catch database generated here matches the nominal catchamounts for each RFMO.

This approach represented the core of the matching methodology,and below we highlight specific refinements that were appliedthroughout the routine to obtain spatialized catch data given availableinformation on country-specific tuna fisheries; all such refinements are

documented in full in the Supplementary Materials. The procedures forthe spatialization of each RFMO’s nominal dataset were developed in Rstatistical programming software (R Core Team, 2016), and are ar-chived in the Sea Around Us GitHub repository (https://github.com/SeaAroundUs/Tuna). We welcome collaborations for further improve-ments.

2.3. Refinements and limitations

As the spatial data reported by RFMOs were limited by the levels ofdetail of reporting by fishing countries and were missing exact matchesfor many records of nominal data as reported by the RFMOs, a series ofrules incorporating additional available information were developed toimprove the quality of the spatialization method. Prior to spatialization,spatial catch records that were erroneously reported as having occurredon land were eliminated from the spatial database to prevent the al-gorithm assigning nominal catch to terrestrial locations. Furthermore,we adjusted suspected but likely accidental spatial misreporting ofSpain fishing in the Atlantic Ocean, as some of the spatial data reportedby Spain were reported only at exact 1° x 1° spatial resolution, but atdistinct 5° intervals. These 1° x 1° reporting records were thus assumedto actually refer to the corresponding 5° x 5° resolution tuna reportingblocks and hence were reassigned to this 5° x 5° resolution. In otherwords, the existing 1° records, which are equally spaced at 5° intervals,were spread over their affiliated 5° x 5° blocks.

Many countries with smaller-scale tuna fishing fleets, defined hereas countries that fish for tuna only within their home FAO area, incontrast to distant-water fleets, do not report gear types to their re-spective RFMO. For instance, Jordan’s tuna fisheries do not report theirgear use to the IOTC, but an independent literature review suggestedthey only use hand lines (Morgan, 2006). Therefore, all of Jordan’s tunacatches were assigned as hand line catches for more accurate spatialassignment. To increase the precision of gear assignment in the spa-tialization, we limited the spatial data used for the matching of nominaldata to spatial data from gears used by a country and excluded spatialdata from gears not used by a country, if so known. This was donethrough a literature review, detailed in the ‘Supplementary Methods:Gear restrictions’ document, where evidence of absence was used as thecriterion to restrict a country to using only specific gear types. For in-stance, French Polynesia has only two tuna fleets: a coastal fleet and anoffshore longline fleet (Misselis, 2003). Therefore, when spatializing thenominal catch of French Polynesia across possible spatial matches whenno gear type is specified in the nominal data being spatialized, thepossible matches are still restricted to the two gear types known to beused in French Polynesia, namely longlines and small-scale gears, ra-ther than all possible gears. In this manner, the spatial extent of afishing country’s nominal tuna catches were restricted to the most likelyactual spatial extent of the relevant gears used.

In addition, the overall spatial extent of some countries’ tuna fish-eries was restricted to a defined area. For example, many countries inthe Red Sea and Persian Gulf area have fisheries that operate morelocally, and thus only within these seas. The lack of matching data inthe spatial database would otherwise spread these catches over a largergeographic zone in the RFMO reporting area (e.g., all of the Western

Table 1Overview of the data sources and tuna block resolution for reporting by each RFMO used here for the development of the harmonized global catch dataset and catchmaps of industrially caught tuna and other large pelagic fishes.

Ocean RFMO Tuna block spatial resolution Countries*/gears/taxa Year of first spatial data

Atlantic ICCAT 1°x1°, 5°x5°, 5°x10°, 10°x10°, 10°x20°, 20°x20° 104/50/164 1950Indian IOTC 1°x1°, 5°x5°, 10°x10°, 10°x20°, 20°x20° 51/35/28 1952Eastern Pacific IATTC 1°x1°, 5°x5° 27/11/21 1954Western Pacific WCPFC 5°x5° 39/10/16 1950Southern CCSBT 5°x5° 9/8/1 1965

* “Countries” includes former countries (i.e. USSR and Yugoslavia) and joint ventures.

A. Coulter, et al. Fisheries Research 221 (2020) 105379

3

Indian Ocean). Restricting these countries’ catches to a smaller areabased on known fishing effort data and tuna fleet characteristics (e.g.,only using dhows) provided a likely more accurate distribution of tunacatches for these countries. See ‘Supplementary Methods: Area restric-tions’ for specific details on this restriction. At the very least, this spatialrestriction reduced the highly unlikely spread of relatively small catchtonnages for a given fishing country over unrealistically large spatialareas. As such area restrictions were only applied to a few fishingcountries in the Indian Ocean (see ‘Supplementary Methods: Area re-strictions’), this restriction had very limited impact on the spatializationof tuna catches.

2.4. Discards

In addition to spatializing all the RFMO-reported nominal landingsdata, we derived preliminary estimates of discards by tuna fisheries byocean basins. Discard rates by fishing country and gear for the PacificOcean were calculated using reconstructed landed and discarded catchof large pelagic species for 1950 to 2010 based on Schiller (2014). Thereconstructed discard rates were projected forward to 2016 by applyingthe average rate from 2005 to 2010. Discard rates were then applied tothe spatialized nominal landings data to calculate discard tonnages forPacific Ocean catches. For industrial fishing fleets, it was assumed thatdiscarding practices would be similar, because the volume and rate ofdiscarded catch is often linked to the value of the species incidentallycaught, and the amount of space onboard at a given point in a trip(Schiller, 2014). In many cases, more coastal smaller-scale fisheries willgenerally target large pelagics and although they have high bycatchrates, much of the non-target catch is retained for local sale or con-sumption rather than discarded at sea. Thus, we assumed no discardingfor the smaller-scale fleets.

A literature review was conducted for the Atlantic and IndianOceans (Le Manach et al., 2016) to collect estimates of discards. Be-cause of the limited amount of country- and fleet-specific data that thissearch revealed, we averaged discard percentages across the entire timeperiod and applied these percentages to the region of origin of a fleet(e.g., Western Europe) rather than the specific country of origin. For theIndian Ocean, our lowest discard percentage was 1.5% (gillnets in Iran,Shahifar, 2012), the median was 7.2% (longline in Asian fleets, IOTC,2000) and the highest percentage was 113% (longline, Alverson et al.,1994). For the Atlantic Ocean, the lowest discard percentage was 1.1%(longlines in North America, ICCAT, 2009), the median was 10.7%(purse-seine in western and northern Europe, Amandè et al., 2011) andthe highest percentage was 100% (swordfish longlines, EuropeanCommission, 2011).

3. Results

3.1. Global

Industrial, or large-scale fisheries for tuna and other large pelagicspecies have expanded globally over the last six decades, as illustratedclearly by the spatial extent of their reported landings data over time(Fig. 2). During the 1950s, catches were largely concentrated off thecoasts of North and Central America in the Eastern Pacific Ocean, andaround the Western Pacific Islands (Fig. 2A). Relatively little fishingeffort, and hence catches, were derived from the Atlantic and IndianOceans. However, by the 1980s, tuna fisheries were active along theequatorial areas across the globe and into temperate regions in mostocean areas (Fig. 2B). In recent years, fisheries catches of tuna andother large pelagic species have been ubiquitous across all tropical andsubtropical regions of the world, with the most catch caught in tropicalareas (Fig. 2C).

Global landed catches as reported by the RFMOs, plus preliminaryestimates of unreported discards of the world-wide tuna fisheries in-creased from around 450,000 t in 1950 to approximately 5.6 million t in

2016 (Fig. 3). Across the whole time period, catches were dominated bythe Pacific Ocean tuna fisheries, which accounted for around 67% oftotal global catches over the full time period, and around 74% in 2016(Fig. 3A). Pacific Ocean tuna catches increased from around 310,000 tin 1950 to over 4.1 million t by 2016 (Fig. 3A). The Atlantic Ocean tunacatches steadily increased from 38,000 t in 1950 to around840,000 t∙year−1 (i.e., 20% of global catch) by the mid-1990s, beforegradually declining thereafter to just over 600,000 t by 2016 (i.e., 12%of global, Fig. 3A). Tuna catches in the Indian Ocean began to grow inimportance only in more recent decades, from around 330,000 t∙year−1

in the mid-1980s (12% of global) to a peak of slightly over990,000 t∙year−1 in the mid-2000s, and around 700,000 t in 2016 (12%of global, Fig. 3A). The Mediterranean Sea has accounted for approxi-mately 2% of the global catch across the study period, growing from21,000 t∙year−1 in the early 1950s to around 100,000 t in 2016(Fig. 3A).

Globally, catches of skipjack tuna have dominated over time, ac-counting on average for 34% of total global catches over the 1950–2016time period (Fig. 3B). Skipjack catches increased from around 103,000 tin 1950 (23% of global catch) to 2.7 million t by 2016 (49% of global).Yellowfin tuna has the second highest catch globally, with catchespeaking in 2003 at 1.5 million t (29% of global), before gradually de-clining thereafter to around 1.4 million t, or 22% of global catches in2016 (Fig. 3B). Other major taxa in the global catch include bigeye tunawith a global average catch of 10% and a generally declining trend toaround 430,000 t in 2016, and albacore tuna, with around 9% of globaltotal catches at around 250,000 t in 2016 (Fig. 3B). In addition to the 12major tuna and billfish taxa detailed in the FAO Tuna Atlas, the globaltuna fisheries currently catch over 700,000 t·year−1 of so-called ‘minortaxa’, or what we here call ‘non-Atlas’ taxa (Fig. 3C). Blue shark(Prionace glauca) has the highest catch of these non-Atlas taxa, withmassive increases in catches over the last two decades to nearly160,000 t in 2016 (Fig. 3C). Other important non-Atlas taxa includeAtlantic bonito (Sarda sarda), little tunny (Euthynnus alletteratus),common dolphinfish (Coryphaena hippurus) and frigate tuna (Auxisthazard).

Overall, reported landings accounted for 89% of total catches, whileour unreported preliminary discard estimates accounted for the re-maining 11% (Fig. 3A, B). However, among the non-Atlas taxa, re-ported landings data accounted for only 40% of total estimated catches,indicating that discards accounted for 60% of total catches of non-Atlastaxa (Fig. 3C).

3.2. Ocean basins

Catches of large pelagic taxa in the Atlantic Ocean have beendominated by Spain across the whole time period (Fig. 4A). Spain’slarge pelagic catches in the Atlantic Ocean peaked at over 210,000 t in1991 and have since been gradually declining to around106,000 t∙year−1 by the mid-2010s, which accounted for around 17%of total Atlantic catches of large pelagics (Fig. 4A). France, the countrywith the second highest total catch, also peaked in the early 1990s, at130,000 t∙year−1 but has since declined more strongly (Fig. 4A). Japanwas the dominant fishing country in the Atlantic Ocean during the1960s, with a peak catch of over 160,000 t in 1965 accounting for wellover 40% of total catches (Fig. 4A). By 2016, however, Japan’s catchesaccounted for a far more modest catch share of around 8% with38,000 t (Fig. 4A). Of note is the considerable increase in large pelagiccatches assigned to the flag of Ghana over the last two decades, whichhas seen its share increase dramatically since the mid-1990s, to accountfor 13% (79,000 t·year−1) of total Atlantic catches by the mid-2010s(Fig. 4A). Given the absence of truly domestic industrial tuna fleets inGhana, these are foreign majority-owned vessels operating under jointventure arrangements, and are dominated by Japanese and Koreanvessels (Nunoo et al., 2014).

During the 1950s, catches in the Atlantic Ocean were mainly

A. Coulter, et al. Fisheries Research 221 (2020) 105379

4

Fig. 2. Global catches of tuna and large pelagic fisheries as spatially assigned by the present study to the variously sized tuna reporting blocks used by each RFMO,averaged for a) 1950–1954; b) 1980–1984; and c) 2012-2016. The appearance of larger and smaller tuna blocks in this figure is an artefact of differences in spatialreporting resolutions from RFMOs, with RFMO spatial reporting blocks ranging from 1°x1° degrees (small dots) to 20°x20° degrees. Most common, and dominatingthe global map are 5°x5° degree tuna reporting blocks (see Table 1).

A. Coulter, et al. Fisheries Research 221 (2020) 105379

5

comprised of Atlantic Bluefin tuna and albacore tuna, while skipjacktuna and yellowfin tuna catches rapidly increased after 1960 (Fig. 4B).More recently, catches are dominated by skipjack (39% or∼230,000 t·year−1), yellowfin (16% or ∼100,000 t·year-1) and bigeyetuna (13% or ∼75,000 t·year-1; Fig. 4B). The fishing gears primarilyused to fish in the Atlantic are longline and purse seine, with a recentstrong increase in purse seine fishing (Fig. 4C).

In the Pacific Ocean, Japan has dominated as the main fishingcountry since 1950, peaking at 900,000 t (40% of total) in 1986, beforedeclining more recently to 10% of total Pacific catches (around350,000 t·year−1; Fig. 5A). The USA is also a major fishing country inPacific waters, at its peak in 1987 accounting for 27% of total catches(540,000 t), while catches declined to around 400,000 t·year−1 by themid-2010s (Fig. 5A). Since the 1990s, other Asian countries have con-tributed significantly to the catch in the Pacific, including Taiwan, In-donesia, South Korea and the Philippines. These four Asian countriesaccounted for nearly 1.5 million t of catch in 2016 (Fig. 5A).

Skipjack and yellowfin tuna are the main taxa caught in the PacificOcean tuna fisheries, together accounting for 76% of total large-pelagiccatch in the Pacific over the full time period, and around 3.0 milliont·year−1 by the mid-2010s (Fig. 5B). Surprisingly (since it is supposedlynot a main target species), the fifth most caught large-pelagic taxon inthe Pacific Ocean is not a tuna taxon, but blue shark, with 82,000 tbeing caught in 2016 (Fig. 5B). The fishing gears that are most wide-spread in the Pacific Ocean tuna fleets are purse seine (69% of the 2016catch) followed by longline (17%; Fig. 5C).

Catches of large pelagics in the Indian Ocean were dominated byJapan until the late-1960s, but by 2016, Japan only accounted for 2%of total catches (20,000 t, Fig. 6A). The massive growth in catches in the1980s and 1990s was driven largely by Taiwan, Spain, France and In-donesia, which together accounted for 44% (around 320,000 t) of total

catches in 2016 (Fig. 6A).Unlike in the Atlantic and Pacific Oceans, yellowfin tuna and skip-

jack tuna were equal contributors over the full time period in the IndianOcean (∼18% each, Fig. 6B), although more recently, catches ofskipjack tuna have been higher (23% in the 2010s). The catch ofSouthern bluefin tuna contributed on average 36% of the catch in the1960s (Fig. 6B). It is now<1% of the catch in the Indian Ocean.Longline gears were used almost exclusively in Indian Ocean tunafisheries up to the late 1960s (Fig. 6C). Subsequently, catching tuna viapurse seine, and to a smaller extent gillnet gears, became popular andnow constitute 55% of the catch in the Indian Ocean (Fig. 6C).

The reported tuna catches in the Mediterranean Sea presented hereshow remarkable, and questionable fluctuations across time, most no-tably due to the biggest fishing countries in the industrial sector, Spainand Turkey (Fig. 7A, see discussion on substantial data reporting issuesfor these two countries). Unlike in the other ocean basins, Mediterra-nean tuna fisheries are predominantly a ‘local’ rather than distant-waterfleet affair, with countries such as Italy, France, Turkey and Spain ac-counting for over 86% of total catches (Fig. 7A).

The taxon with highest cumulative catch in the tuna fisheries in theMediterranean was Atlantic bonito, accounting for 25% of total timeseries catch, but more recently only 22% in 2016 (Fig. 7B). Atlanticbluefin tuna and swordfish comprise two other historically importanttaxa in these waters (Fig. 7B). However, 44% of the total catch acrossthe time period were non-Atlas taxa, including blue shark and Atlanticbonito, which also heavily dominated total catches in the last decade(Fig. 7B). Gear use varied widely over time, with longline and purseseine dominating in recent times, while gillnet, pole-and-line and traps(the last two gears are pooled into ‘other’) were widely used in earlierdecades (Fig. 7C).

Fig. 3. Global catches of tuna and other large pelagic fishes from 1950 to 2016as assembled and harmonized from the five separate tuna RFMO datasets, by a)ocean basins; b) major taxa (156 additional taxa are pooled in ‘Other’); and c)important taxa beyond the 12 major target species covered in the FAO Atlas ofTuna and Billfish Catches (144 additional taxa are pooled in ‘Other’).

Fig. 4. Catches of tuna and large pelagics in the Atlantic Ocean from 1950 to2016, by a) fishing country (107 additional countries are pooled in ‘Other’); b)taxon (134 additional taxa are pooled in ‘Other’); and c) fishing gear (9 addi-tional gear categories are pooled in ‘Other’).

A. Coulter, et al. Fisheries Research 221 (2020) 105379

6

4. Discussion

Here we present a globally harmonized and spatially assigned da-taset of catches (reported landings, plus unreported preliminary discardestimates) of large pelagic fishes integrated across all ocean basins for1950–2016, and include tunas, billfishes, sharks and smaller scombrids.Several separate, individual spatial datasets on large pelagic fisheriescatches exist elsewhere, either for subsets of catches by ocean basinsreported by RFMOs and globally (Taconet et al., 2017), or globally forthe 12 major tuna and billfish taxa as part of the FAO Tuna Atlas (FAO,2017). However, the present study is the first attempt at harmonizingall datasets across ocean basins, and assigning all nominal catches pluspreliminary estimates of discards to spatial tuna data reporting blocksof 1° x 1° to 20° x 20° spatial resolution.

Tuna fisheries, here referring to ‘all large pelagics’ including bill-fishes, sharks and other large- and medium-sized scombrids, illustratethe greatest spatial expansion in global fisheries (Swartz et al., 2010)due to the highly migratory nature of tuna and other large-pelagic taxaand their wide-spread occurrence in high seas waters, i.e., areas beyondnational jurisdiction. However, this expansion is also indicative of therequirement to fish farther offshore as coastal and near-shore areasbecame increasingly depleted. Today, industrial fishing fleet effortcovers at least 55% (Kroodsma et al., 2018) and possibly up to 90%(Tickler et al., 2018b) of the global ocean, and fishing in waters beyondnational jurisdiction (i.e., the high seas) is becoming increasingly un-economical without extensive government subsidies and often ques-tionable cost-cutting labor practices (Sala et al., 2018; Tickler et al.,2018a, b). The scale of this pattern of expansion is troubling, as we havereached the spatial limit for tuna fisheries, with no new fishing groundsremaining, except as provided through species distributional changesdue to climate change, which, however, may also lead to loss of pre-vious fishing grounds and catch potential (Cheung et al., 2009, 2010).

Thus, the continuation of tuna fisheries and their associated economicbenefits at levels similar to the present or recent time periods is de-pendent on effective and restrictive long-term sustainable managementof the fisheries and fleets exploiting these stocks and ecosystems. Giventhe predominance of a small number of highly developed, wealthycountries as the main (Sumaila et al., 2015) as well as highly subsidizedfishing countries (Sumaila et al., 2014, 2016) in the global tuna andlarge pelagic fisheries, such effective actions are achievable withminimal socio-economic impacts, while increasing global incomeequality (Sumaila et al., 2015).

4.1. By-catch and discards

With the exception of swordfish, the stock status of many other non-tuna large pelagic taxa is unknown or uncertain (Majkowski, 2007).Industrial tuna fisheries are one of the major threats to pelagic sharkpopulations (Gilman, 2011), which is worsened by the fundamentallydifferent life history dynamics of sharks (Dulvy et al., 2008). Whilesharks only contributed 5% of the total overall catch presented here,they constituted 12% of the catch caught in tuna longline fisheries.Other studies have also shown that shark longline bycatch is very high(Worm et al., 2013), possibly as high as 25% (Gilman, 2011). Of thecatch of sharks reported to the RFMOs, 66% is blue shark. This is po-tentially worrisome due to this species’ very low resilience and an IUCNRed List Status of ‘near threatened’ (Stevens, 2009). Over 6 million t ofshark landings were reported to the RFMOs since 1950 and these wereoften reported in generic pooled taxonomic groupings such as Elas-mobranchii and Selachimorpha, which are highly uninformative taxo-nomic groupings. Our estimate of taxon-specific discards for the PacificOcean alone suggests a further 5.7 million t of sharks were discardedacross the time period. Despite contributing a large portion of the catchin tuna fisheries, sharks are often not recorded to the species level

Fig. 5. Catches of tuna and large pelagics in the Pacific Ocean from 1950 to2016, by a) fishing country (55 additional countries are pooled in ‘Other’); b)taxon (38 additional taxa are pooled in ‘Other’); and c) fishing gear (6 addi-tional gear categories are pooled in ‘Other’).

Fig. 6. Catches of tuna and large pelagics in the Indian Ocean from 1950 to2016, by a) fishing country (54 additional countries are pooled in ‘Other’); b)taxon (83 additional taxa are pooled in ‘Other’); and c) fishing gear (5 addi-tional gear categories are pooled in ‘Other’).

A. Coulter, et al. Fisheries Research 221 (2020) 105379

7

(Dulvy et al., 2008), which makes managing fishing pressure and con-serving threatened species extremely difficult (Dulvy et al., 2000). Allof the RFMOs, except the CCSBT, currently have mitigation measures toreduce non-target large pelagic bycatch, such as temporal closures toreduce capture of undersized and non-target tunas, and restrictions onshark finning practices.

However, none of the RFMOs set precautionary catch or bycatchlimits for sharks, nor do they require better taxonomic resolution ofdata (Gilman, 2011), and only the WCPFC requires gear technologies toreduce the capture of sharks (WCPFC, 2010a). Though widely proposedas a mitigation measure, modifications to fishing strategies needagreement between tuna fishing countries and enforcement policies forsuccess (Poisson et al., 2016).

The practice of discarding is wide-spread in commercial fisheries,even though there is agreement that it is both ecologically and eco-nomically undesirable (Alverson et al., 1994; Kelleher, 2005; Zelleret al., 2018; Zeller and Pauly, 2005). Discarding involves a consciousdecision made by fishers to abandon a portion of their catch for reg-ulatory or economic reasons (Bellido et al., 2011). While discardedcatch is often thought of as primarily composed of non-target, un-marketable taxa, including pelagic sharks (Gilman, 2011), there ismuch evidence that it also includes juveniles of otherwise marketablespecies (Kelleher, 2005), or species valuable to other fishing sectorssuch as artisanal fisheries. Just as importantly, ‘high-grading’, i.e., se-lective retention of relatively higher-value fish at the expense of pre-vious catch, is very common in tuna fisheries, where vessels prefer tooptimize hold space over the course of long voyages (Hall et al., 2000).

To mitigate discarding, fisheries management entities increasinglymandate the landing of all target catch, though this is difficult to im-plement unless mandatory and independent 100% observer coverage isenforced. Furthermore, this does not address discarding of non-targetcatch. While the use of at-sea observers for monitoring is costly, this

should be considered part of normal operational costs for sustainablefisheries. Observers may be pressured by crew to violate standards toavoid confrontation (Robertson, 1998), which would require immediateresponses by RFMOs and the responsible flag state of the offendingfishing vessel(s). Observer coverage varies between RFMOs and geartypes, with a minimum of 5% agreed upon across RFMOs (WCPFC,2010b). This extremely low level of coverage is a clear and criticaldeficit of RFMO governance (Gilman, 2011; WCPFC, 2010b). As tech-nological innovations such as on-board camera coverage and satellitemonitoring are now widely available for fisheries monitoring, an in-dustrial tuna fishery with 100% unbiased coverage is readily feasible(McCauley et al., 2016) and should be foundational to any fisheriesaiming for sustainability and accountability (Zeller et al., 2011). Pilotprojects to implement such technological tools for tuna fisheries havebeen undertaken, with a positive response from industry members forquite some time (e.g., Piasente et al., 2012). Thus, substantial and in-tense efforts should be expended by all RFMOs and responsible membercountries to implement and enforce such measures for all fleets.

Our estimates of discards for tuna fisheries as presented here shouldbe treated as preliminary in nature and may in many cases under-estimate the scale of the wasteful discard problem. The discard ratesused here only account for a subset of the literature, and difficultiesexist in harmonizing them. Furthermore, IATTC and ICCAT both reportsome discards in their databases, but the extent of these data are un-clear. It appears that FAO may have erroneously counted the discardsreported in the ICCAT database as landings in the FAO dataset, as thetotal landed tonnage reported in FAO’s Global Capture Productionstatistics is equal to the sum of all discards and landings in the ICCATdatabase (FAO, 2019; ICCAT, 2016). This contradicts FAO’s stated datareporting objective to not include discarded catches in their GlobalCapture Production database. A full and detailed reconstruction ofdiscards for the Atlantic and Indian Oceans, as well as an update andrevision of this work for the Pacific Ocean (Schiller, 2014), carefullyintegrating RFMO reports of discards and scientific literature of discardtonnage and rates, would better capture the global extent of discardingin tuna fisheries. Feedback, input and offers of collaboration from ex-perts is welcome, and will allow us to refine these rates. We welcomesuch collaborations to improve the quality of global data.

4.2. Comparison to other global analyses

The FAO compiles RFMO data into the FAO FishStat global capturedatabase (FAO, 2019), and reports on the global spatial catch of a se-lected subset of tuna in their Tuna Atlas (FAO, 2017). Here, our ap-proach builds on these efforts by harmonizing all RFMO datasets and byspatially assigning all nominal reported catch data. Currently, the FAOTuna Atlas only covers 12 target species of high value tuna and bill-fishes. While this encompasses much of the economically importantcatch by global tuna fisheries, it only conveys a partial view of the truescope of the industry. All catch in the FAO Tuna Atlas were also ag-gregated into 5° x 5° blocks, though much of this information is actuallyavailable at a resolution of 1° x 1°. This loss of spatial resolution isproblematic in today’s context of fisheries management’s need to ac-count for ecosystem considerations (Juan-Jordá et al., 2018; Pikitchet al., 2004), and in our view should be avoided.

There have been other excellent efforts to compile global databasesfor tuna fisheries, many of which contributed to the FAO Tuna Atlas.Fonteneau (1997) published a global atlas, but did not estimate dis-cards, nor scaled up the spatialized data to 100% of the nominal catch.Updates were published later, but at regional scales and without thePacific Ocean (Fonteneau, 2009, 2010). The recent efforts by the IRDand FAO on harmonizing tuna RFMO data into a single database is akinto our ideal spatialization scenario with matching flag, species, yearand gear combinations (Taconet et al., 2017). Unfortunately, just likeFonteneau (1997), they did not scale up the spatialized data to 100% ofthe nominal catch, as was done in the present study, despite obviously

Fig. 7. Catches of tuna and large pelagics in the Mediterranean Sea from 1950to 2016, by a) fishing country (61 additional countries are pooled in ‘Other’); b)taxon (103 additional taxa are pooled in ‘Other’); and c) fishing gear (10 ad-ditional gear categories are pooled in ‘Other’).

A. Coulter, et al. Fisheries Research 221 (2020) 105379

8

higher data uncertainty, using a set of logical assumptions in the ab-sence of partial parameter match. Finally, the International SeafoodSustainability Foundation (ISSF, 2017) has produced biannual technicalreports of the status of the world fisheries for tuna by compiling RFMOdata, though they only address the 23 major tuna stocks on the scale ofocean basin (ISSF, 2019).

The spatial data used for our analysis were sourced as public datafrom the various tuna RFMOs, and we compliment the RFMOs and FAOon the challenging job of annually assembling these data, which cor-relate well with AIS-derived fishing effort data by the Global FishingWatch (Kroodsma et al., 2018). Still, the officially reported datasets aresubject to misreporting and cover a limited set of the tuna fisheriesmanaged by the RFMOs. Some spatial misreporting was obvious, e.g.,when tuna catches were reported at spatial coordinates that occur onland. Public datasets are also subject to confidentiality rules of theRFMOs, which unfortunately leads to incomplete or even entirelymissing data components. Most notably, data from the IATTC had alarge region of the North Pacific Ocean almost totally devoid of longlinefishing data until the early 1990s. Publicly releasing data for this region(even if in aggregated form to mask the identities of fishing enterprises)would facilitate investigations into the spatial extent of historical tunafisheries and improve transparency and accountability around the useof what is actually a public resource. Besides, hiding data from publicview or use does not make sense these days, give the wide publicavailability of additional data sources, e.g., the satellite-derived data onspatial fishing effort provided by the Global Fishing Watch (Kroodsmaet al., 2018). Moreover, vague and qualitative geographic information(i.e., ‘sub-areas’) such as provided in the ICCAT nominal catch databasewas not usable as these ‘sub-areas’ are not explicitly defined geo-graphically.

The present harmonization was limited by the data reported to theRFMOs. For regions where underreporting is known to occur regularly,for example Spain and Turkey fishing in the Mediterranean Sea (seeFig. 7 and Pauly et al., 2014b), full catch reconstructions (Zeller et al.,2016) are required to complement the incomplete reported data withbest estimates of unreported activities to comprehensively account forthe extent of global tuna fisheries. Furthermore, the role of China intuna fisheries will have to be re-assessed, given that its heavily sub-sidized distant-water fleets, including those targeting tuna, are gradu-ally muscling out those of other distant-water fishing countries (Paulyet al., 2014a). Therefore, future data efforts by tuna RFMOs shouldemphasize both more comprehensive reporting by all fishing fleets, aswell as more detailed spatial reporting efforts by countries.

Improving the state of global tuna stocks is key for sustaining highlyvaluable fisheries that are particularly important for Small IslandDeveloping States as well as other tropical and subtropical developingcountries that have historically benefitted from these species.Transparency and full accountability are vital components for propermanagement, particularly in the context of supporting these regions, asspecifically included in the UN Sustainable Development Goals (SDG14.7), and more broadly for ensuring a more equitable distribution ofbenefits from these highly migratory species (Sumaila et al., 2015).

CRediT authorship contribution statement

Angie Coulter: Conceptualization, Methodology, Software, Formalanalysis, Investigation, Writing - original draft, Writing - review &editing, Visualization. Tim Cashion: Conceptualization, Methodology,Software, Formal analysis, Investigation, Writing - original draft,Writing - review & editing. Andrés M. Cisneros-Montemayor:Methodology, Software, Writing - review & editing. Sarah Popov:Software, Writing - review & editing, Visualization. Gordon Tsui:Software, Writing - review & editing. Frédéric Le Manach:Methodology, Software, Writing - review & editing. Laurenne Schiller:Writing - review & editing. Maria Lourdes D. Palomares:Conceptualization, Writing - review & editing, Supervision, Project

administration, Funding acquisition. Dirk Zeller: Conceptualization,Writing - review & editing, Supervision, Project administration,Funding acquisition. Daniel Pauly: Conceptualization, Writing - review& editing, Supervision, Project administration, Funding acquisition.

Acknowledgements

The authors acknowledge the support of the Sea Around Us and theSea Around Us – Indian Ocean at the University of British Columbia andthe University of Western Australia, respectively. All Sea Around Usactivities are supported by the Oak Foundation, the Marisla Foundation,Oceana, the MAVA Foundation, the David and Lucile PackardFoundation, Bloomberg Philanthropies, the Paul M. Angell FamilyFoundation and the Minderoo Foundation. None of the funders had anyinput into the data, analysis, interpretation or writing of this manu-script.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in theonline version, at doi:https://doi.org/10.1016/j.fishres.2019.105379.

References

Alverson, D.L., Freeberg, M.H., Pope, J.G., Murawski, S.A., 1994. A Global Assessment ofFisheries by-catch and Discards. FAO Fisheries Technical Papers T339, Food andAgriculture Organization (FAO), Rome.

Amandè, M.J., Ariz, J., Chassot, E., Chavance, P., Delgado de Molina, A., Gaertner, D.,Murua, H., Pianet, R., Ruiz, J., 2011. By-catch and discards of the European purseseine Tuna fishery in the Atlantic Ocean: estimation and characteristics for 2008 and2009. Collective Volume Scientific Papers - ICCAT 65, 2113–2120.

Bellido, J.M., Santos, M.B., Pennino, M.G., Valeiras, X., Pierce, G.J., 2011. Fishery dis-cards and bycatch: solutions for an ecosystem approach to fisheries management?Hydrobiologia 670, 317.

Cheung, W., Lam, V., Sarmiento, J., Kearney, K., Watson, R., Pauly, D., 2009. Projectingglobal marine biodiversity impacts under climate change scenarios. Fish Fish. 10,235–251.

Cheung, W.W.L., Lam, V., Sarmiento, J., Kearney, K., Watson, R., Zeller, D., Pauly, D.,2010. Large-scale redistribution of maximum fisheries catch potential in the globalocean under climate change. Glob. Change Biol. 16, 24–35.

Cullis-Suzuki, S., Pauly, D., 2010. Failing the high seas: a global evaluation of regionalfisheries management organizations. Mar. Policy 34, 1036–1042.

Cullis-Suzuki, S., Pauly, D., 2016. Global evaluation of high seas fisheries management.In: Paul, D., Zeller, D. (Eds.), Global Atlas of Marine Fisheries: A Critical Appraisal ofCatches and Ecosystem Impacts. Island Press, Washington D.C, pp. 79–85.

Dulvy, N.K., Metcalfe, J.D., Glanville, J., Pawson, M.G., Reynolds, J.D., 2000. Fisherystability, local extinctions, and shifts in community structure in skates. Conserv. Biol.14, 283–293.

Dulvy, N.K., Baum, J.K., Clarke, S., Compagno, L.J.V., Cortés, E., Domingo, A., Fordham,S., Fowler, S., Francis, M.P., Gibson, C., Martínez, J., Musick, J.A., Soldo, A., Stevens,J.D., Valenti, S., 2008. You can swim but you can’t hide: the global status and con-servation of oceanic pelagic sharks and rays. Aquat Conserv: Mar Freshwat Ecosyst.18, 459–482.

European Commission, 2011. Impact Assessment of Discard Reducing Policies - EUDiscard Annex. Studies in the Field of the Common Fisheries Policy and MaritimeAffairs - Lot 4: Impact Assessment Studies Related to the CFP. European Commission.

FAO, 2016. The State of World Fisheries and Aquaculture 2016 (SOFIA): Contributing toFood Security and Nutrition for All. Food and Agriculture Organization, Rome.

FAO, 2017. Atlas of Tuna and Billfish Catches (www document). Fishery StatisticalCollections. [online: www.fao.org/fishery/statistics/tuna-atlas/en Last accessed: 23March 2017]. Food and Agriculture Organization, Rome.

FAO, 2018. The State of World Fisheries and Aquaculture 2018 (SOFIA): Meeting theSustainable Development Goals. Food and Agriculture Organization, Rome.

FAO, 2019. FishStatJ - Software for Fishery and Aquaculture Statistical Time Series.V3.5.3. Food and Agriculture Organization, Rome. [online www.fao.org/fishery/statistics/software/fishstatj/en].

Fonteneau, A., 1997. Atlas of Tropical Tuna Fisheries - World Catches and Environment.Office de la Recherche Scientifique et Technique Outre-Mer (ORSTOM), Paris(France).

Fonteneau, A., 2009. Atlas of Atlantic Ocean Tuna Fisheries. IRD Editions, Paris (France),pp. 170.

Fonteneau, A., 2010. Atlas of Indian Ocean Tuna Fisheries. IRD Editions, Paris (France),pp. 189.

Garibaldi, L., 2012. The FAO global capture production database: a six-decade effort tocatch the trend. Mar. Policy 36, 760–768.

Gillett, R., 2007. A Short History of Industrial Fishing in the Pacific Islands. Asia-PacificFishery Commission and Food and Agriculture Organization of the United Nations(FAO), Bangkok (Thailand).

A. Coulter, et al. Fisheries Research 221 (2020) 105379

9

Gilman, E.L., 2011. Bycatch governance and best practice mitigation technology in globaltuna fisheries. Mar. Policy 35, 590–609.

Hall, M.A., Alverson, D.L., Metuzals, K.I., 2000. By-catch: problems and solutions. Mar.Pollut. Bull. 41, 204–219.

ICCAT, 2009. Report of the 2008 ICCAT Yellowfin and Skipjack Stock AssessmentsMeeting, July 21-29 2008, Florianópolis (Brazil). SCRS/2008/016. Collective Volumeof Scientific Papers. ICCAT. 64, pp. 669–927.

ICCAT, 2016. Rules and Procedures for the Protection, Access and Dissemination of DataCompiled by ICCAT. International Commission for the Conservation of Atlantic Tuna.Nominal Catch Information Database. ICCAT report 2010-2011 (I) Annex 6.

IOTC, 2000. Indian Ocean Tuna Fisheries - Data Summary for 1989-1998. IOTC DataSummary 20. Indian Ocean Tuna Commission (IOTC), Victoria (Seychelles).

Status of the World Fisheries for Tuna. ISSF, Washington, D.C. USA.ISSF, 2019. Status of the World Fisheries for Tuna: March 2019. International Seafood

Sustainability Foundation, ISSF Technical Report 2019-07, Washington D.C.Juan-Jordá, M.J., Murua, H., Arrizabalaga, H., Dulvy, N.K., Restrepo, V., 2018. Report

card on ecosystem-based fisheries management in tuna regional fisheries manage-ment organizations. Fish Fish. 19, 321–339.

Kelleher, K., 2005. Discards in the World’s Marine Fisheries. An Update. FAO FisheriesTechnical Paper 470, Food and Agriculture Organization, Rome.

Kroodsma, D.A., Mayorga, J., Hochberg, T., Miller, N.A., Boerder, K., Ferretti, F., Wilson,A., Bergman, B., White, T.D., Block, B.A., Woods, P., Sullivan, B., Costello, C., Worm,B., 2018. Tracking the global footprint of fisheries. Science 359, 904–908.

Lam, V.W.Y., Sumaila, U.R., Dyck, A., Pauly, D., Watson, R., 2011. Construction and firstapplications of a global cost of fishing database. ICES J. Mar. Sci. 68, 1996–2004.

Le Manach, F., Chavance, P., Cisneros-Montemayor, A.M., Lindop, A., Padilla, A., Schiller,L., Zeller, D., Pauly, D., 2016. Global catches of large pelagic fishes, with emphasis onthe High seas. In: Pauly, D., Zeller, D. (Eds.), Global Atlas of Marine Fisheries: ACritical Appraisal of Catches and Ecosystem Impacts. Island Press, Washington, D.C,pp. 34–45.

Majkowski, J., 2007. Global Fishery Resources of Tuna and Tuna-like Species. FAOFisheries Technical Paper No. 483. Food and Agriculture Organization, Rome.

McCauley, D.J., Woods, P., Sullivan, B., Bergman, B., Jablonicky, C., Roan, A., Hirshfield,M., Boerder, K., Worm, B., 2016. Ending hide and seek at sea. Science 351,1148–1150.

Misselis, C., 2003. French Polynesia tuna fisheries. In: National Tuna Fishery ReportPresented to the 15th Meeting of the Standing Committee on Tuna and Billfish(Honolulu, Hawaii, 18-27 July, 2002). Honolulu.

Miyake, M.P., Miyabe, N., Nakano, H., 2004. Historical Trends of Tuna Catches in theWorld. FAO Fisheries Technical Paper. No. 467. Food and Agriculture Organization,Rome.

Morgan, G., 2006. Country review: Jordan. In: De Young, C. (Ed.), Review of the State ofWorld Marine Capture Fisheries Management. Indian Ocean. FAO Fisheries TechnicalPaper No. 488. Food and Agriculture Organization, Rome, pp. 255–260.

Nunoo, F.K.E., Asiedu, B., Amador, K., Belhabib, D., Pauly, D., 2014. Reconstruction ofMarine Fisheries Catches for Ghana, 1950-2010. Fisheries Centre Working Paper#2014-13, Fisheries Centre. University of British Columbia, Vancouver.

O’Connor, S., Ono, R., Clarkson, C., 2011. Pelagic fishing at 42,000 years before thepresent and the maritime skills of modern humans. Science 334, 1117–1121.

Pauly, D., Belhabib, D., Blomeyer, R., Cheung, W.W.L., Cisneros-Montemayor, A.,Copeland, D., Harper, S., Lam, V., Mai, Y., Le Manach, F., Österblom, H., Mok, K.M.,van der Meer, L., Sanz, A., Shon, S., Sumaila, U.R., Swartz, W., Watson, R., Zhai, Y.,Zeller, D., 2014a. China’s distant water fisheries in the 21st century. Fish Fish. 15,474–488.

Pauly, D., Ulman, A., Piroddi, C., Bultel, E., Coll, M., 2014b. ‘Reported’ versus ‘likely’fisheries catches of four Mediterranean countries. Scienta Marina. 78, 11–17.

Piasente, M., Stanley, B., Timmiss, T., McElderry, H., Pria, M., Dyas, M., 2012. ElectronicOnboard Monitoring Pilot Project for the Eastern Tuna and Billfish Fishery.Australian Fisheries Management Authority. FRDC Project 2009/048, Canberra.

Pikitch, E.K., Santora, C., Babcock, E.A., Bakun, A., Bonfil, R., Conover, D.O., Dayton, P.,Doukakis, P., Fluharty, D.L., Heneman, B., Houde, E.D., Link, J., Livingston, P.A.,Mangel, M., McAllister, M.K., Pope, J., Sainsbury, K.J., 2004. Ecosystem-basedfishery management. Science 305, 346–347.

Poisson, F., Crespo, F.A., Ellis, J.R., Chavance, P., Pascal, B., Santos, M.N., Séret, B., Korta,M., Coelho, R., Ariz, J., Murua, H., 2016. Technical mitigation measures for sharksand rays in fisheries for tuna and tuna-like species: turning possibility into reality.Aquat. Living Resour. 29, 402.

R Core Team, 2016. R: a Language and Environment for Statistical Computing. RFoundation for Statistical Computing, Vienna, Austria.

Robertson, G., 1998. The culture and practice of longline tuna fishing: implications forseabird by-catch mitigation. Bird Conserv. Int. 8, 211–221.

Sala, E., Mayorga, J., Costello, C., Kroodsma, D., Palomares, M.L.D., Pauly, D., Sumaila,U.R., Zeller, D., 2018. The economics of fishing the high seas. Sci. Adv. 4, eaat2504.

Schiller, L., 2014. Tuna Be, or Not Tuna Be: Using Catch Data to Observe the EcologicalImpacts of Commercial Tuna Fisheries in the Pacific Ocean at Varying Spatial Scales.MSc thesis Available at:. University of British Columbia, Vancouver, pp. 155. https://open.library.ubc.ca/cIRcle/collections/ubctheses/24/items/1.0166022.

Schiller, L., Bailey, M., Jacquet, J., Sala, E., 2018. High seas fisheries play a negligible rolein addressing global food security. Sci. Adv. 4, eaat8351.

Shahifar, R., 2012. Estimation of bycatch and discard in Iranian fishing vessels (gillnets)in the IOTC area of competence during 2012. In: 8Th Session of the IOTC WorkingParty on Ecosystems and Bycatch, IOTC-2012-WPEB08-42, Indian Ocean TunaCommission (IOTC. Cape Town, South Africa.

Stevens, J., 2009. Prionace glauca. The IUCN Red List of Threatened Species 2009e.T39381A10222811. https://doi.org/10.2305/IUCN.UK.2009-2.RLTS.T39381A10222811.en. Downloaded on 04 September 2019. .

Sumaila, U.R., Dyck, A.J., Baske, A., 2014. Subsidies to tuna fisheries in the WesternCentral Pacific Ocean. Mar. Policy 43, 288–294.

Sumaila, U.R., Lam, V.W.Y., Miller, D., Teh, L., Watson, R., Zeller, D., Cheung, W.W.L.,Côté, I.M., Rogers, A., Roberts, C., Sala, E., Pauly, D., 2015. Winners and losers in aworld where the high seas is closed to fishing. Nature Scientific Reports. 5, 8481.

Sumaila, U.R., Lam, V., Le Manach, F., Swartz, W., Pauly, D., 2016. Global fisheriessubsidies: an updated estimate. Mar. Policy 69, 189–193.

Swartz, W., Sala, E., Tracey, S., Watson, R., Pauly, D., 2010. The spatial expansion andecological footprint of fisheries (1950 to present). PLoS One 5, e15143.

Taconet, P., Chassot, E., Guitton, J., Palma, C., Fiorellato, F., Anello, E., Barde, J., 2017.Global database and common toolbox for tuna fisheies. Collective Volume ScientificPapers ICCAT. 73, 3327–3337.

Tickler, D., Meeuwig, J.J., Bryant, K., David, F., Forrest, J.A.H., Gordon, E., Larsen, J.J.,Oh, B., Pauly, D., Sumaila, U.R., Zeller, D., 2018a. Modern slavery and the race tofish. Nat. Commun. 9, 4643.

Tickler, D., Meeuwig, J.J., Palomares, M.-L., Pauly, D., Zeller, D., 2018b. Far from home:distance patterns of global fishing fleets. Sci. Adv. 4, eaar3279.

WCPFC, 2010a. Conservation and Management Measure for Sharks. Western and CentralPacific Fisheries Commission Seventh Regular Session. Conservation andManagement Measure 2010-071, Honolulu. Available at: https://www.wcpfc.int/doc/cmm-2010-07/conservation-and-management-measure-sharks [Accessed: 04-September-2019].

WCPFC, 2010b. Report of the International Workshop on Tuna RFMO Management ofIssues Relating to Bycatch. Western and Central Pacific Fisheries CommissionScientific Committee Sixth Regular session. WCPFC-SC6-2010/EB- IP-05, Nuku’alofa,Tonga. Available at: https://www.wcpfc.int/node/2504 [Accessed: 01-September-2019].

Worm, B., Davis, B., Kettemer, L., Ward-Paige, C.A., Chapman, D., Heithaus, M.R., Kessel,S.T., Gruber, S.H., 2013. Global catches, exploitation rates, and rebuilding options forsharks. Mar. Policy 40, 194–204.

Zeller, D., Pauly, D., 2005. Good news, bad news: global fisheries discards are declining,but so are total catches. Fish Fish. 6, 156–159.

Zeller, D., Rossing, P., Harper, S., Persson, L., Booth, S., Pauly, D., 2011. The Baltic Sea:estimates of total fisheries removals 1950-2007. Fish. Res. 108, 356–363.

Zeller, D., Palomares, M.L.D., Tavakolie, A., Ang, M., Belhabib, D., Cheung, W.W.L., Lam,V.W.Y., Sy, E., Tsui, G., Zylich, K., Pauly, D., 2016. Still catching attention: SeaAround Us reconstructed global catch data, their spatial expression and public ac-cessibility. Mar. Policy 70, 145–152.

Zeller, D., Cashion, T., Palomares, M.L.D., Pauly, D., 2018. Global marine fisheries dis-cards: a synthesis of reconstructed data. Fish Fish. 19, 30–39.

A. Coulter, et al. Fisheries Research 221 (2020) 105379

10